Chemistry of bicyclo [3.1. 0] hexatriene

Journal of the American Chemical Society. Advanced Search. Search; Citation; Subject. Search in: Anywhere, Title, Author, Abstract. J. Am. Chem. Soc...
0 downloads 0 Views 416KB Size
7828 Scheme I

0"' t

electrocyclic ring opening converting the diene into the product triene. IV would also be the precursor of the 6-tert-butoxyfulvene 111. As shown in Scheme I, path a, the transformation could entail a second HBr elimination generating the triene I followed by addition of t-BuOH and then a similar electrocyclic opening. Several alternative pathways would convert IV to I11 without necessitating further loss of HBr. Grohmann had suggested nucleophilic attack of tert-butoxide a t c6 with concomitant ring opening and loss of bromide (path b).2 Addition-elimination to 6-bromofulvene would also form 111 (path c). Invoking Scheme I, one can predict the result of replacement of a bridgehead hydrogen with deuterium. For example, the primary isotope effect of a deuterium a t Cl would enhance formation of bromobenzene a t the expense of both fulvenes. However, if the deuterium were a t C5, IV deuterated a t C5 should be formed at a rate comparable to the undeuterated material. The partitioning of IV between 6-bromo and 6tert-butoxyfulvene should be unchanged if both were the result of only one HBr elimination. However, if 111arose via I, the amount of 6-tert-butoxyfulvene should decrease due to the primary isotope effect. To test this analysis, we prepared the monodeuterated dibromide 11, containing one deuterium equally distributed between the bridgehead carbons, from 6-deuterio- l ,4-dioxa-6spiro[4.4]nonene ( V ) which was prepared by bromination, dehydrobromination, and Na/t-BuOD reduction of 1,4dioxa-6-spiro[4.4] nonene. V was converted to I by sequential reactions with CHzBrz/LiN(SiMe3)2, H 3 0 + , SeO2, AlH3, and HBr.3 Under conditions identical with those employed for the undeuterated dibromide, bridgehead deuterated I1 was converted in poor yield to 6-tert-butoxyfulvene; the major products were bromobenzene and 6-bromofulvene. These results demonstrate that the loss of both bridgehead hydrogens was required for the formation of the 6-tert-butoxyfulvene. This observation is consistent with a mechanism entailing the intermediacy of bicyclo[3.1.O]hexatrieneI and not a pathway entailing only one HBr elimination. The following procedure was used to analyze for deuterium content and location. The standard conditions for the transformation of deuterated I1 to 6-dimethylaminofulvene (VI) were as follows: 1 m M of I1 in 1 ml of THF was added under argon over 30 min to 3 ml of THF at -75 OC containing 6 m M of KO-t-Bu and 8 m M of HNMe2. After stirring for 1 h at -75 OC, the solution was quenched by H2O. T o prevent any possibility of exchange, the aqueous washes were kept neutral or a l k a l i ~ ~The e . ~ deuterium content of the product aminofulvene VI was determined by mass spectral analysis after conversion to 6 - p h e n y l f ~ l v e n eTo . ~ ~ascertain ~ the location of the deuterium, we utilized the observation that the hydrogens a t Cl and C4 of VI are easily exchanged by mild acid catalysis.' Moreover, under more vigorous conditions, exchange will even occur a t C 2 and C3. Typically, the product VI was split into two portions: only one of which was exchanged. Both samples were then converted to the 6-phenylfulvene. Comparison of the deuterium content of each established the location of the label as being distributed a t one or more of the three distinct sets of positions: c6, C I and Cq, and C2 and C3. We could not distinguish between Cl and C4 or between C2 and C3 by this method. Consistent with formation of I, the mixture of bridgehead Journal of the American Chemical Society

monodeuterated dibromides was converted in poor yield to VI containing a total of 0.09 D (0.03 D a t C2 and C3 and 0.06 D a t C1 and C4). To exclude the possibility of hidden skeletal rearrangements, we prepared I1 deuterated at C3 and at cf,.I1 deuterated a t C3 was prepared by acid catalyzed exchange of the a-methylene protons of 6-bromo-2-bicyclo[3.1 .O] hexanone followed by the procedure stated above. The dibromide I1 containing 0.5 D at C3 was converted in excellent yield to VI containing 0.32 D a t CZand C3 and 0.06 D a t C I and Cq. To label c6, I1 was prepared from deuterated benzvalene.8 The labeled dibromide contained 1.5 total D of which 0.75 was a t c,5,0.22 a t c2 and c4, respectively, and 0.15 at C I and c g , respectively. Upon conversion in high yield to the aminofulvene VI, followed by exchange, 0.7 D remained in the molecule. By IH NMR, all the deuterium was a t c6. From these two labeling studies it is apparent that no skeletal rearrangements occurred. Moreover, deuterium a t c3 or c6, or as will be seen at C2 or Cq, had no effect on the reaction c o ~ r s e Only . ~ deuterium a t either bridgehead position adversely affected the formation of 6-tert-butoxy or 6-dimethylaminofulvene. These results are consistent only with the formation of bicyclo[3.1 .O]hexatriene, the chemistry of which is described in the following c o m m u n i ~ a t i o n . ~

Acknowledgments. W e thank Research Corporation for financial support. References and Notes (1) W. Washburn, J. Am. Chem. SOC., 97, 1615 (1975). (2) A. H. Amaro and K. Grohmann, J. Am. Chem. SOC.,97, 5946 (1975). (3) The 3:l exo/endo mixture of epimers of Cg was separated as the enones. (4) Separate control experiments with VI dideuterated at C, and C4 and Cp and CBconfirmed that no deuterium loss occurred either under the reaction or workup conditions. (5) Mass spectral analysis of VI to ascertain deuterium content was unreliable since controls revealed appreciable loss of deuterium from the ring carbons. (6) 6-Phenylfulvenes were prepared by addition of phenyilithium to VI followed by dilute acetic acid. Control experiments established that deuterium could not be washed in or out under these conditions. (7) A. Mannschreck and U. Kolle, Chem. Ber., 102, 243 (1969). (8) T. J. Katz and N. Acton, J. Am. Chem. Soc., 95, 2738 (1973). (9) W. Washburn and R. Zahler, J. Am. Chem. Soc., following paper in this issue.

William N. Washburn,* Robert Zahler Department of Chemistry, Unicersity of California. Berkeley Berkeley, California 94720 Received May 21, 1976

Chemistry of Bicyclo[3.1 .O]hexatriene Sir: This paper focuses on two aspects of the chemistry of bicyclo[3.1.0]hexatriene (I): ( 1 ) the acidity of I and (2) the transformation of I to fulvenes. The complex nature of the latter was revealed by the examination of the products arising from a mixture of 2- and 4-deuterio-exo,exo-4,6-dibromobicyclo[3.1 .O] hexenes (11) prepared by alane-d3 reduction of 6-bromo-2-bicyclo[3.1 .O]hexenone followed by exposure to HBr. By IH N M R the sole deuterium was equally divided between C2 and C4. Under the standard conditions previously outlined,' this labeled dibromide was converted to 6-dimethylaminofulvene (111) containing 0.3 D distributed between C I and C4.2 The amount of deuterium washed out varied between 0.6 and 0.8 D, depending on the reaction conditions. The S N ~products ' isolated from the reaction had not lost deuterium as judged by ' H N M R , thus precluding prior exchange of 11.

/ 98:24 / November 24, 1976

7829

I n contrast, when a 1:l mixture of [2-13C]- and [4-I3C]4,6-dibromobicyclo[3.1 .O]hexene, prepared from [ 1-I3C] cyclopentanone as previously described,' was converted to 111, the I3C label was equally distributed over all four carbons: C I , Cz, C3, and C4. The apparent incompatability of these two results prompted computational investigations using GAUSSIAN 704 in conjunction with the STO-3G5 basis set. The isodesmic reaction (eq 1 ) was employed to estimate the basicity of the centers of high electron density, i.e., C2 and C4, by combining our calculations with respect to I with similar studies entailing the protonation of substituted benzene.6 I was predicted to be more basic than N,N-dimethylaniline. In view of the basicity of Cz and C4 we suggest that the initial addition of HNMez to I occurs via nucleophilic attack a t c6 accompanied by simultaneous proton delivery to either site to form 6-dimethylaminobicyclo[3.1 .O]hexa-1 I,j,2-diene (IV). A [ 1,5] hydrogen shift would convert IV to 6-dimethylaminobicyc10[3.1.O]hexa1,3-diene ( V ) which would be transformed to the product fulvene by a disrotatory electrocyclic opening.

I

X

To explore whether the two possible [ 1,5] transition states were of similar energy, we compared the S C F energy for each transition state with the hydrogen midway between the two carbons involved. W e found a substantial preference (12 kcal) for the migration of the hydrogen cis to the nitrogen. To our knowledge this is the first instance of u asymmetry so perturbing a symmetrical system that only one of two otherwise equivalent [ 1,5]sigmatropic hydrogen migrations would be expected to occur. One rationale for this event is as follows. Of the two u bonds at c6 the carbon-hydrogen bond is of higher energy and thus should mix preferentially with the P system. Applying Frontier orbital concepts invoking orbital distortion, one would predict higher electron density at C I and Cs on the molecular face anti to this carbon- hydrogen u b ~ n dFrom . ~ ~examination ~ of the interaction of this P system with the two hydrogen atoms (A and B), preferential migration of the hydrogen which can maintain maximum bonding in the transition state (A) is not surprising.

Scheme I

I

P

I *&Me2

xu

to Cq would result in formation of 4-deuterio-6-dimethylaminofulvene. Delivery of the proton to Cz of the deuterated triene would generatea 4-deuterio-6-dimethylaminobicyclo[3.1 .O]hexa-1 '.j,2-diene which must be eventually transformed to a 2-deuteriodimethylaminofulvene. For this to occur, the deuterium must have exclusively migrated since failure of the deuterium to migrate would have resulted in its appearance at C3 of the product fulvene. In this instance, the electronic stabilization of the transition state should more than compensate for the unfavorable kH/kD isotope effect.I0 Furthermore, since the u asymmetry at Cb enhances only the migratory apptitude of the atom cis to the nitrogen, the deuterium must be cis, implying a trans addition of dimethylamine to the triene I. In view of the T electron polarization of I, one might expect the c6 hydrogen to be even more acidic than cyclopropene. We sought evidence, both theoretical and experimental, for the enhanced acidity of I. Streitwieser and McKelveyI2," had found a linear relationship between experimental pK,'s of a hydrogen bound to an sp' carbon and the energy difference between the hydrocarbons and the corresponding anion obtained by G A U S S I A N 70 calculations. From consideration of the isodesmic reaction (eq 2) where R H was benzene, ethylene, and cyclopropene, we predict that I should be more acidic than any of the above.I3 Given a pK, of 43 for benzene,14 the theoretically estimated pK, of I is -40-41. When I1 was converted to I11 as previously described, except that 15 equiv of DNMez was present, 0.2 D was incorporated a t c6 and -0.03 D on the ring of III.I5 The amount of exchange at c6 could be increased to 0.7 D by adding 1 equiv of 18-crown-6-ether to the reaction medium. No effect was observed on deuterium incorporation of the ring. Moreover, due to the increased base strength, a second reaction pathway producing bromobenzene became competitive. Analysis of exo,exo- 2-dimethylamino-6-bromobicyclo[3.1 .O] hexene formed by competitive Sy2' attack on 11 confirmed that prior exchange of I1 had not occurred at c6 under these conditions. Likewise, controls established that the c6 hydrogen of 111 was not labile. We cannot exclude the possibility of exchange occurring from the product resulting from an initial 1,4 elimination of HBr, Le., exo-6-bromobicyclo[3.1.0] hexa-1,3-diene (VI). However, we found that the amount of deuterium incorporation at c6 under identical conditions was 0.4 D for I1 and 0.43 D for the isomeric exo,endo-2,6-dibromobicyclo[3.1 .O]hex2-ene. Since the overlap of the 7 system with the u bond of the syn and anti hydrogens differs substantially, one would have expected more of an effect if exchange had occurred at this point.

gf WNMe2 NMe2

A

B

Applying these considerations to the two products expected from nucleophilic addition of HNMe2 to the resonance stabilized Cz labeled triene I leads to the following predictions (Scheme I): For I3C labeled I, protonation a t C2 followed by [1,5] migration and ring opening would result in the label appearing at Cz of the product fulvene. In contrast, protonation a t C4 would ultimately generate a C4 labeled fulvene. The nucleophilic medium interconverts C I and Cq and C2 and C3 by addition/elimination to c6 of IIL9 Thus the I3Clabel should appear a t all ring carbons a and 0to the exocyclic methylene as was observed. For the Cz deuterated triene, proton delivery

Communications to the Editor

7830 The failure to observe substantial amounts of deuterium in the five-membered ring of I11 would indicate that the t-BuOH involved in the nucleophilic addition to I must contain one of the original bridgehead protons. A substantial primary isotope effect in conjunction with a solvent cage during the protonation of I would account for preferential hydrogen incorporation into 111. To date, we cannot completely factor the contribution due to solvent cage and isotope effects. However, the net result is that very little of the bulk deuterium of the solvent appears in 111.

The fact that varying amounts of exchange of deuterium originally present at C2 and C3 of I1 occurred suggests that the eliminations are not concerted. Presumably, the intermediate cyclopentadienyl anion VI1 leading from VI to I can be reversably protonated on carbons other than the bridgehead, thus providing a mechanism for deuterium loss. Once again, a primary isotope effect would account for the failure to observe appreciable deuterium incorporation. Cram has commented on a similar experience in which deuterium was lost more easily to a protio solvent pool than a hydrogen to a deuterio one involving the base promoted isomerization of I-protio- and 1deuterio- 1-methyl-3-tert- butylindenes. l 6 The conversion of IT to 111 is summarized in Scheme I. Since labeling studies revealed loss of deuterium from CZand Ci, the anion VI1 must be reversably protonated a t all ring positions. However, to maintain the stereochemical integrity discussed above, I once formed is converted to 111 in concerted reactions. I f an anion of IV or V were formed, nonspecific protonation would obligate the C Zdeuterated triene to be converted to 111 which would contain deuterium a t all four ring carbons. Moreover, the rearrangement of IV to V entails an unique instance of the u asymmetry a t Cg distinguishing between two otherwise equivalent electrocyclic reactions.

Acknowledgment. W e thank Research Corporation for financial support. W e thank Dr. John McKelvey for assistance with the estimates of the acidity of I. References and Notes W. Washburn and R. Zahler. J. Amer. Chem. SOC.,preceding paper in this issue. By 100-MHz ’H NMR in CH3CN3deuterium appeared to be both at C1and C .,. However, the low amount of label and partial overlap of the C4and CS protons precluded rigorous assignment of amounts at each carbon. For ’H NMR assignment of the ring protons of 111 see A. Mannschreck and U.Kolle. Chem. Ber., 102, 243 (1969). W. J. Hehre, W. A. Lathan, R. Ditchfield, M. D. Newton, and J. A. Pople, Quantum Chemistry Program Exchange. University of Indiana. Blwmington, Ind., Program No. 236. See W. J. Hehre, R. F. Stewart, and J. A. Pople, J. Chem. Phys., 51, 2657 (1969). J. M. McKelvey, S. Alexandratos, A. Streitwieser, J. M. Abboud, and W. J. Hehre, J. Am. Chem. SOC.,98, 244 (1976). K. Fukui, Acc. Chem. Res., 4, 57 (1971). C. L. Liotta, TetrahedronLett., 519, 523 (1975). This process proceeded at a measurable rate for tert-butyl-ring substituted 6-dimethylaminofulvene but would be expected to be much faster for Ill since the alkyl inductive effect would destabilize the intermediate anion. For example the pKa of tert-butylcyclopentadiene is -19 compared to a pK, of 16 for unsubstituted cyclopentadiene. (R. Zahler unpublished work.) At 25 O C typical values are 1-1.4 kcal.” C. W. Spangler, Chem. Rev., 76, 187 (1976). A. Streitwieser and J. M. McKelvey, unpublished work. The STO-3G basis set was employed for all calculations. A. Streitwieser, P. J. Scannon. and H. M. Niemeyer, J. Am. Chem. Soc., 94, 7936 (1972). As discussed previously,’ a significant loss of deuterium did not occur in the conversion of C6 deuterated dibromide iI to the aminofulvene Ill reflecting a primary isotope effect. In the presence of 18crown-6ether 70% deuterium loss was observed. J. Almy, D. H. Hoffman, K. C. Chu, and D. J. Cram, J. Am. Chem. SOC.,95, 1185 (1973).

William N. Washburn,* Robert Zahler Department of Chemistry, University of California, Berkeley Berkeley, California 94720 Receiued M a y 21, I976

Journal of the American Chemical Society / 98:24

Use of Circular Dichroism Measurements to Determine the Formation Constant of the Complex between an Optically Active Benzo-15 -crown-5 and Sodium Ion Sir:

Circular dichroism (CD) provides a powerful tool with which to study the conformation of molecules in solution.’ An obviously essential condition is that the substrate under study be optically active. To facilitate interpretation of the C D data, the structure of the compound should be as simple as possible, and incorporate at least one chromophore which absorbs in an accessible region of the electronic spectrum, thereby exhibiting a measurable Cotton effect. Although conformational freedom makes unambiguous interpretation of C D data difficult in most cases, the number of possible conformers is conveniently restricted in the case of cyclic molecules. One subset of these, the cyclic polydentate molecules,2 is especially interesting in that additional restrictions are imposed on the conformation of the molecule when it binds a cation. Blout3 has recently reviewed much of the research on the cyclic peptide ionophores in this connection. In our laboratory attention has been directed toward those macrocyclic polyethers termed “crowns” by their discoverer Pedersen.? They possess the advantage of a high degree of solvolytic stability due to the inertness of the ether bond, and can readily be designed with considerable symmetry, and to incorporate at least one aromatic ring. This in turn provides two useful absorption bands in the ultraviolet, each potentially leading to a Cotton effect in optically active derivatives. It is also helpful that the literature provides information about the preparation and characterization of a number of the parent (achiral) compound^.^ In this communication we wish to report the synthesis of an optically active “1 5-crown-5”, specifically (8S,12S)-8,12-dimethyl-5,6,8,9,11,12,14,15-octahydro[b] benzo-l,4,7,10,13-pentaoxacyclopentadecin(VII), and the use of circular dichroism measurements to calculate the equilibrium constant governing the formation of its complex with sodium ion (as the perchlorate) in spectrograde methanol at ambient temperature. The polyether VI1 was synthesized as described in Scheme I; compound V is commercially available,6 while the synthesis of I has been previously reported.’ The elemental analysis of VII, its proton N M R , IR, and UV spectra were all consistent with the assigned structure. Its mass spectrum exhibited a strong parent ion peak a t 296 mass units. Figure 1 shows the C D spectra of a 1.51 X M solution of VI1 in methanol, alone and in the presence of a 2000-fold molar (formal) excess of NaC104. (There were no significant differences between the latter curve, and those obtained with

-4-1

I

I

u

p 4 1 1 x

-8

a, -16 -I2!

--

210 2 2 0 2 3 0

260 270

280

Figure 1. (A) C D spectrum of a 1.51 X IO-4 M solution of VI1 i n anhydrous methanol, ambient temperature; JASCO O R D / C D / 5 , SS-20 modification. (B) Same as A except with addition of 250- to 2000-fold excess NaC104.

/ November 24, 1976